# Estimating Power from MelCloud data **URL:** \ **Category:** Heatpump\ **Tags:** melcloud, ecodan, emontx, emoncms\ **Created:** [25 September 2024 13:34 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767 "2024-09-25T13:34:54Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![abbbbbottttt](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/a/c57346/32.png) [@abbbbbottttt](https://community.openenergymonitor.org/u/abbbbbottttt)\ **Post date:** [25 September 2024 13:34 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/1 "2024-09-25T13:34:54Z")
Hi, I’m looking to take my heat pump monitoring a step further and have been playing with the MyHeatPump app. My current set up involves feeds of data from emonTx4 (power only), MelCloud, and Octopus. This gets me pretty close to everything I need to feed into the app. I’m hoping to be able to fill in the blanks - the most prominent of which is heat output. - I fully understand that this figure is not going to be 100% accurate due to MelCloud but want to get as close as I can without adding to the setup. \* MelCloud provides Flow and Return. I know the flow rate to be 24l/min. I have been looking at the calculations and believe this should get me to a power figure which I can accumulate to kWh. Is this a reasonable approach? Is processing the inputs the easiest way to achieve this? Or can I do anything with virtual feeds? Do I need to take anything else into account? Pumps, OperationMode, etc? Many Thanks, Chris ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [25 September 2024 15:07 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/2 "2024-09-25T15:07:04Z")
I have been doing something similar to estimate heat from Melcloud for a handful of systems. I’ve calculated the heat output for different flow rates. 24 l/min equates to 1672 W per K of dT, where dT is the difference between flow and return. If 20% glycol then this drops to 1572. - water: 4182 x _flow rate_ / 60 - glycol: 4000 x _flow rate_ / 60 (If you’re able to read actual flow rate, use that instead) The feeds you need to pull from MelCloud are: - **FlowTemperature** - **ReturnTemperature** - **WaterPump1Status** - primary pump - **ValveStatus3Way** - 0 = heating, 1 = dhw Then use input processing to compute the following: **FlowTemperature** - ReturnTemperature × WaterPump1Status × WperK → heat and heat\_kwh ![image](https://community.openenergymonitor.org/uploads/default/original/3X/d/7/d7e6315163ef3b7117fd7390779303e237ca160f.png) As this uses `Log` and `Power to kWh` processes, so can’t be done with virtual feeds. I also have a method to estimate electrical consumption from MelCloud, but you don’t need that if you have power from emonTX. ---
**Author:** ![abbbbbottttt](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/a/c57346/32.png) [@abbbbbottttt](https://community.openenergymonitor.org/u/abbbbbottttt)\ **Post date:** [25 September 2024 17:03 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/3 "2024-09-25T17:03:34Z")
@Timbones - Thank you so much for the detailed response. Do you use a similar calculation for the hot water? And will the MyHeatPump app then sort out COP, etc? ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [25 September 2024 17:54 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/4 "2024-09-25T17:54:26Z")
Hot water may have a different flow rate, but I’ve not tried to compensate for that. Difficult to fiddle with using input steps, but could be calculated separately and then combined. Yes, these will plot nicely in My Heatpump app. Here an example compared against carnot factor 0.5: ![Screenshot_20240925_185129_com_android_chrome_ChromeTabbedActivity](https://community.openenergymonitor.org/uploads/default/original/3X/2/e/2e1fd5d7c09036dafdc57e89e939a6d044f4e00d.jpeg) Use WaterPump2Status for central heating signal. ---
**Author:** ![abbbbbottttt](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/a/c57346/32.png) [@abbbbbottttt](https://community.openenergymonitor.org/u/abbbbbottttt)\ **Post date:** [25 September 2024 22:19 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/5 "2024-09-25T22:19:28Z")
Thank you. Really appreciate your help. ---
**Author:** ![F1p](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/f/b77776/32.png) [@F1p](https://community.openenergymonitor.org/u/F1p)\ **Post date:** [26 September 2024 05:26 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/6 "2024-09-26T05:26:25Z")
Hi @Timbones, I was wondering about the input and output power from MELCloud as at the CN105, Input power seems to be: 0 if between 0 and 1kW, 1 if between 1 and 2kW, Etc. I would be interested to see your formula for input power estimation Output power seems to be rounded to the nearest kW, but both in & out reported on the CN105 to 0d.p. ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 06:56 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/7 "2024-09-26T06:56:44Z")
When fitted with meters\*, Melcloud reports watt-hours for the past minute. Multiply by 60 to get current power in watts. \*CanMeasureEnergyConsumed value will be true. When there’s no meter, input power can be estimated using **HeatPumpFrequency** , which is compressor ~~percentage~~ Hz. Divide by 100 and multiply by the maximum electric input for that model in watts. I.e. 3,270 for 8.5 kW model, 3,730 for the 11.2 kW model. I also add in the pumps to give something a little more realistic, but it’s very crude. Oh, plus 20W for standby. ![Screenshot_20240926_074900_com_android_chrome_ChromeTabbedActivity](https://community.openenergymonitor.org/uploads/default/original/3X/c/3/c3ef64f380d88e998b012ad7309eebfd286f41a7.jpeg) _This formula has been superceded by a more accurate one described in [post #19](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/19)_ 👇 ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 07:01 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/8 "2024-09-26T07:01:10Z")
> [@abbbbbottttt](#): > > I know the flow rate to be 24l/min. For anyone who needs to know: Flow rate can be read from Running Information on the FTC unit, code `540`. It’s not available from MelCloud. **Beware** that glycol systems will show a slightly lower flow rate by a couple l/min, so will need to adjust for that. ---
**Author:** ![abbbbbottttt](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/a/c57346/32.png) [@abbbbbottttt](https://community.openenergymonitor.org/u/abbbbbottttt)\ **Post date:** [26 September 2024 08:04 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/9 "2024-09-26T08:04:22Z")
> [@Timbones](#): > > **Beware** that glycol systems will show a slightly lower flow rate by a couple l/min, so will need to adjust for that. @Timbones - So what is available in FTC may not be accurate? What sort of magnitude of adjustment are we talking? ---
**Author:** ![F1p](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/f/b77776/32.png) [@F1p](https://community.openenergymonitor.org/u/F1p)\ **Post date:** [26 September 2024 08:30 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/10 "2024-09-26T08:30:00Z")
Two points of clarification please - > [@Timbones](#): > > input power can be estimated using **HeatPumpFrequency** , which is compressor percentage The MELCloud value is not in Hz when this is the units presented by the CN105? > [@Timbones](#): > > Divide by a hundred and multiply by the maximum electric input for that model in watts How do you account for frequency limiting in Eco hot water (60 or 64Hz maximum)? ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 09:03 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/11 "2024-09-26T09:03:52Z")
> [@F1p](#): > > The MELCloud value is not in Hz when this is the units presented by the CN105? I dunno, maybe. For the systems I’ve looked at this happens to map to 100 Hz = full power, and matches the readings I get from my electric meter. Lucky coincidence I guess. ![image](https://community.openenergymonitor.org/uploads/default/original/3X/d/b/dbafcf652d2602ef7537595d4033348cb1ea7355.png) - Measured vs. Estimated electricity on left-hand axis up to max input watts for my unit. - HeatPumpFrequency (red) on right-hand axis up to 100 Hz. I have struggled to sensible numbers from larger units, so maybe they go higher than 100 Hz? So knowing that maximum would allow converting Hz to %. I’ll dig into the history for those units to see if there’s a correlation… > [@F1p](#): > > How do you account for frequency limiting in Eco hot water (60 or 64Hz maximum)? I don’t see why this would matter. This is just equates to less power use, right? ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 09:10 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/12 "2024-09-26T09:10:49Z")
> [@abbbbbottttt](#): > > So what is available in FTC may not be accurate? What sort of magnitude of adjustment are we talking? Umm… may need to add 15-25%? FTC claims 12 or 13 l/min for my system, yet calculating the flow from the measured heat & dT comes closer to 15. I presume because the Sika is not calibrated for glycol. I’ve not tried reading the actual flow rate from the heat meter as it’s hard to get to. ---
**Author:** ![F1p](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/f/b77776/32.png) [@F1p](https://community.openenergymonitor.org/u/F1p)\ **Post date:** [26 September 2024 09:43 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/13 "2024-09-26T09:43:33Z")
> [@Timbones](#): > > I.e. 3,270 for 8 kW model, 3,730 for the 11.2 kW model. Based on my rating sticker: ![image](https://community.openenergymonitor.org/uploads/default/original/3X/4/0/40566dce50b5d5da6fc1cd987e0b8ee02ba8b4d0.jpeg) Is maximum 2668W or 2093W? > [@Timbones](#): > > 100 Hz = full power, Yeah and this is where i would differ, frequency at Eco cap of 64Hz and consumption continues up to 2670W. Then swap to heating, frequency up to 78Hz and input power down to ~2000W ![image](https://community.openenergymonitor.org/uploads/default/original/3X/a/c/ac378337a06bde598bdce72276b9b323bf3d335d.png) ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 10:16 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/14 "2024-09-26T10:16:05Z")
> [@F1p](#): > > Based on my rating sticker: > > Is maximum 2668W or 2093W? Looking at the [manufacturer’s spec sheet](https://library.mitsubishielectric.co.uk/pdf/download_full/4145) [PDF], maximum input is 3.27 kW. ![image](https://community.openenergymonitor.org/uploads/default/original/3X/d/6/d6850492efc48cfbbc5d1c7148ff5dca8e066ca2.png) > [@F1p](#): > > Yeah and this is where i would differ, frequency at Eco cap of 64Hz and consumption continues up to 2670W. > Then swap to heating, frequency up to 78Hz and input power down to ~2000W I guess this is due to the compressor having to work harder as the refrigerant gets hotter, and so pulls more power for the same frequency. I see the same during the hot water cycle. (34 Hz, not Eco, but Quiet Mode 1). ![image](https://community.openenergymonitor.org/uploads/default/original/3X/4/e/4eb47e45644152da78b34856325af5b563a847a0.png) So, this estimation of electrical consumption isn’t great and will produce optimistic COP during DHW runs. Could maybe include an extra fiddle factor for flow temperature, I guess. It does seem to be good enough for diagnosing; a couple users have been able to improve performance of their systems based on this data. ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [26 September 2024 10:59 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/15 "2024-09-26T10:59:46Z")
> [@Timbones](#): > > I’ve pre-calculated the heat output for different flow rates. 24 l/min equates to 1672 W per K of dT, where dT is the difference between flow and return. If 20% glycol then this drops to 1572. This is based on equations for [mass flow rate heat transfer](https://docs.openenergymonitor.org/heatpumps/basics.html#mass-flow-rate-heat-transfer). * * * An alternative approach uses **[Carnot COP equation](https://docs.openenergymonitor.org/heatpumps/basics.html#carnot-cop-equation)**, which might work better where an electric meter is installed. Multiply measured electrical power by Carnot COP to get an estimate of heating power. Can view this directly in the My Heatpump app, using a typical factor around **0.4** - **0.5**. Useful for sanity checking the other estimation method. ![image](https://community.openenergymonitor.org/uploads/default/original/3X/8/a/8a89913cfdc642b771f451d677956cc5ea50a2e5.png) Or calculate it programmatically. I use this for a Vaillant system that comes with MID approved electric meter, but doesn’t report flow rate or return temperature so can’t use mass flow equations: ```python # Carnot heat calculation condensing = flowT + 2 + 273 evaporating = outdoorT - 6 + 273 carnot_cop = condensing / (condensing - evaporating) * 0.5 if electric < 100: # ignore standby power carnot_heat = 0 carnot_heat = electric * carnot_cop ``` Note the 100W threshold for ignoring standby + pump power; this may need to be higher for larger systems (also a setting in MyHP app). It would be more accurate to _subtract_ the standby + pump power rather than cutting off at the threshold. Or have separate meters for compressor, pumps and controller. ---
**Author:** ![F1p](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/f/b77776/32.png) [@F1p](https://community.openenergymonitor.org/u/F1p)\ **Post date:** [26 September 2024 12:44 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/16 "2024-09-26T12:44:34Z")
> [@Timbones](#): > > Could maybe include an extra fiddle factor for flow temperature, I guess. This is how i was thinking of doing it - or, trying to determine based on onboard monitoring where it might be: > [@F1p](#): > > 0 if between 0 and 1kW, > 1 if between 1 and 2kW, ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [16 December 2024 20:29 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/17 "2024-12-16T20:29:55Z")
Did you figure out a better formula for estimating power draw? I’m taking on a **5kW** system to monitor remotely, and would like to get the numbers better if I can. ---
**Author:** ![F1p](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/f/b77776/32.png) [@F1p](https://community.openenergymonitor.org/u/F1p)\ **Post date:** [16 December 2024 21:33 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/18 "2024-12-16T21:33:23Z")
Well reports are that pretty good for 8.5kW, as expected it’s slightly over estimating for a 5kW and slightly under estimate for a 11 or 14kW. It’s reasonable enough to make big changes, but probably not to fine tune. I’m not really sure how to compensate for the different sizes as there isn’t anything in the FTC to suggest the outdoor unit size without a user input… ---
**Author:** ![Timbones](https://community.openenergymonitor.org/user_avatar/community.openenergymonitor.org/timbones/32/20308_2.png) [@Timbones](https://community.openenergymonitor.org/u/Timbones)\ **Post date:** [18 December 2024 18:04 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/19 "2024-12-18T18:04:46Z")
I’ve spent some time figuring out a better way for estimating how much power the heatpump draws. I started with the performance data from _Ecodan\_ATW\_Databook\_R32\_Vol5.3\_.pdf_ that shows the Max\* Capacity and COP: (\*“Performance at Maximum compressor frequency”) ![image](https://community.openenergymonitor.org/uploads/default/original/3X/5/a/5a2dfdeef64d3f873181a8cbe7a285273b102e1b.png) Capacity ÷ COP gives me the input power in kW. Plotted as watts against Flow Temp - Ambient Temp is very nearly linear. I ignored values below -10° and above 15° as these didn’t fit well and aren’t typical for UK climate. I also ignored 2° ambient as that includes defrosts, and omitted 25° flow. ![image](https://community.openenergymonitor.org/uploads/default/original/3X/4/8/488b32cd1580cb1fbf2e9a03a5ee33143b611423.png) The trend lines fit with R² of 0.9, and give us formulas for each of the three models I looked at: (I did look at fitting a quadratic instead, but the end result wasn’t much different) - **5 kW** - `26` dT + `530` - **8.5 kW** - `54.7` dT + `1062` - **11 kW** - `68` dT + `976` From this I was able to work out the following formula implemented in a virtual feed: 1. Flow temp - Outside temp 2. × factor for model, e.g. `68` 3. + offset for model, e.g. `976` 4. × HeatPumpFrequency ÷ maximum frequency¹ 5. + 250 W if compressor is running² 6. + 100 W if primary pump is running³ 7. + 20 W for standby power ¹ max frequency possibly different for each model, I don’t know. 110 Hz for my system ² I guess this is for the big fan and other things going on when compressor is running ³ pump power probably depends on the speed setting… Here’s measured vs. estimated for my system on Dec 14th: ![image](https://community.openenergymonitor.org/uploads/default/original/3X/a/d/ad0788ea92c4ab738d7b9a84a081abf7978bc26f.png) Yellow is the estimated power, red is the measured power for heating, blue for hot water. I’ve had to add an extra 400W for the DHW pump, which seems excessive but is what fits. Same formula running on another 11.2 kW system shows it matching just as well there too: ![image](https://community.openenergymonitor.org/uploads/default/original/3X/4/5/4531825edd511c4ae4b5cc4a75969d186c6694b1.png) ---
**Author:** ![toadhall](https://community.openenergymonitor.org/letter_avatar_proxy/v4/letter/t/34f0e0/32.png) [@toadhall](https://community.openenergymonitor.org/u/toadhall)\ **Post date:** [19 December 2024 09:42 UTC](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767/20 "2024-12-19T09:42:37Z")
> [@Timbones](#): > > I presume because the Sika is not calibrated for glycol. It should work without recalibration. From the SIKA web site … Alternate vortices rotating in opposite directions are generated behind a bluff body immersed in a flow. The vortices detach from the edges of the bluff body and form a Kármán vortex street in the fluid stream. The distance between the single vortices is constant. The frequency of the vortices flowing past a sensor depends on the flow rate and is proportional to the flow. The sensor detects these vortices which are then converted to an electrical frequency signal. Elsewhere on the web … - Unaffected by temperature, pressure, density, or viscosity of the media Theodore von Karman, a Hungarian-American physicist, was the first to describe the effect where a non-streamlined object (also called a bluff body) placed in the path of a fast-flowing stream, causes the fluid to alternately separate from the object on its two downstream sides, and, as the boundary layer becomes detached and curls back on itself, forming vortices (also called whirlpools or eddies). He also noted that the distance between the vortices was constant and depended solely on the size of the rock that formed it. [Next page](https://community.openenergymonitor.org/t/estimating-power-from-melcloud-data/26767.md?page=2)